基于多元监测信息融合的含裂隙类岩石材料破坏规律和预警方法

Failure law and early warning method of precast fractured rock specimen based on multi monitoring information fusion

为探究含裂隙岩石试件的变形破裂特征和前兆规律,进行了含裂隙类岩石材料的破坏试验,采用数字图像相关技术(DIC)、红外辐射测温技术(IRT)和声发射技术(AE)实时无损地探测了试件在损伤破坏过程中应变场、红外辐射温度场及声发射振铃的响应规律、前兆性和敏感性,并引入样本标准差指标对应变场和温度场矩阵数据进行了数据降维和定量分析。基于上述物理量监测信息,建立了考虑多元监测信息及其权重的数据融合函数。在此基础上,对试件破坏过程进行了阶段划分和综合预警。结果表明:试件在裂隙起裂、扩展和贯通过程中表现出显著的应变局部化特征。试件破坏过程中,应变局部化带逐渐扩大,应变场标准差曲线呈非线性变化。试件破坏过程中产生的多物理量响应规律和前兆性具有一定区别。红外辐射温度场和声发射振铃次数在宏观裂隙生成前变化平稳,在宏观裂隙生成时发生突变。而应变场在裂隙起裂前就展现出应变离异特征,前兆性更显著。多元监测信息反应试件破坏的显著性也不相同,强弱顺序为水平应变场标准差、剪切应变场标准差、垂直应变场标准差、温度场标准差、声发射振铃次数。综合考虑各指标特点,建立了基于多元信息共生和各监测信息权重的综合评判方法,并从能量积累的角度对试件的破坏过程进行量化分析,确定了任意时刻试件宏观破坏发生的累积概率,实现了分级预警和概率预警。

In order to explore the deformation and fracture mechanism and precursory characteristics of rock specimens with fractures, the failure tests of rock materials with fractures were carried out. The damage process of the specimen, the strain field, infrared radiation temperature field and the response law and precursory of acoustic emission ringing were detected non destructively in real time by using digital image correlation technology(DIC), infrared radiation temperature measurement technology(IRT) and acoustic emission technology(AE). The sample standard deviation index is introduced to analyze the data dimensionality reduction and quantitative analysis of the matrix data of strain field and temperature field. As a result, the stage division and comprehensive early warning of the specimen failure process are carried out. The results show that: The specimen exhibits significant strain localization characteristics during crack initiation, propagation and coalescence. In the process of specimen failure, strain localization band gradually expands, and the curve of standard deviation of strain field shows nonlinear change. The response law of multi physical quantities produced in the failure process of the specimen is different from the precursor. The infrared radiation temperature field and acoustic emission ringing times change steadily before the formation of macro cracks, and change suddenly during the formation of macro cracks. But the strain field shows the characteristics of strain segregation before the crack initiation, and the precursor is more sensitive. Multivariate monitoring information reflects different significance of specimen failure. The sensitive order is the standard deviation of horizontal strain field, the standard deviation of shear strain field, the standard deviation of vertical strain field, the standard deviation of temperature field and the number of acoustic emission rings. Considering the characteristics of each index, a comprehensive evaluation method based on multiple information symbiosis and the weight of each monitoring information is established. The failure process of the specimen is quantitatively analyzed from the perspective of energy accumulation, the cumulative probability of macro failure of the specimen at any time is determined, and the hierarchical early warning and probability early warning are realized.